Backtest v9 — Performance Audit Jan 2021 → Dec 2024 · 4 years 207 closed trades
Tested Performance

+122.8%
vs SPY's +68.3% & QQQ's +69.5%

$10,000 turned into $22,282.61 across a full market cycle — the 2021 bull run, the brutal 2022 bear market, and the 2023–24 recovery. All benchmarks beaten by +54 percentage points.

See the Chart
🏆 WINNER
StockSentry
222.83
+122.8%
$10k → $22,282
SPY
168.27
+68.3%
$10k → $16,827
−54.5pp vs SS
QQQ
169.46
+69.5%
$10k → $16,946
−53.3pp vs SS
Sharpe
1.25
Risk/Reward
~5:1
Win Rate
35.7%
Max DD
−23.7%
Equity Curve vs Benchmarks

Four years. Two benchmarks. One clear winner.

$10,000 starting capital, marked to market monthly. The orange line is StockSentry — indexed to 100. Bear market band shows the Mar–Nov 2022 regime.

StockSentry · +122.8% SPY · +68.3% QQQ · +69.5% Bear Regime
Hover for values
2021 · Bull
+42.9%

StockSentry outperformed SPY (+30.5%) and QQQ (+29.2%) in the momentum phase, riding 10 conviction setups.

2022 · Bear
−13.7%
StockSentry
vs
−32.6%
QQQ

Capital-protection shields cut losses nearly in half vs QQQ — preserving compounding power for the recovery.

2023–24 · Recovery
+35.6% / +36.7%

Rotation engine deployed capital into GE (+142%), META (+67%), MU (+61%), CEG (+34%) — beating SPY both years.

Annual Return Breakdown

Every year, every number.

No cherry-picked window. The key story: the 2022 bear market protection left StockSentry with far more capital to compound in 2023–24.

Year
StockSentry
SPY
QQQ
2021
+42.9%
+30.5%
+29.2%
2022 Bear
−13.7%
BEST IN CLASS ✓
−18.2%
−32.6%
−19pp vs StockSentry
2023
+36.7%
Beat SPY by +10pp
+26.7%
+55.9%
AI/tech bounce year
2024
+35.6%
+25.6%
+27.7%
Total
+122.8%
$10K → $22,283
+68.3%
$10K → $16,827
+69.5%
$10K → $16,946

* Annual returns calculated from first to last trading day of each calendar year. Returns are simulation results, not live trading.

The Asymmetry Engine

Why a 35.7% win rate is the point.

The Sentry doesn't need to be right most of the time. It needs to be rewarded massively when right, and disciplined when wrong.

Average Winning Trade
+20.00%

ATR trailing stop locked profits as momentum compounded. Winners held an average of 130 days.

win size distribution
Average Losing Trade
−4.30%

Circuit-breaker stops fire fast. The system exits before a small loss becomes a painful one. Average hold on losers: 22 days.

loss size distribution
The math

Average win is 4.65× larger than the average loss. Even at a 35.7% win rate, every 100 trades generates a net positive expectancy of +4.83% per trade.

Risk/Reward Ratio
~5:1
win per unit risked
Trade Statistics

Every stat. No spin.

207 closed trades over 4 years. Here are the raw numbers — good and bad.

Final Equity
$22,282.61

From $10,000 starting capital

Closed Trades
207

74 winners · 133 losers · small avg loss

Sharpe Ratio
1.25

Risk-adjusted return efficiency

Max Drawdown
−23.7%

Peak-to-trough. QQQ hit −46% intra-period.

Win Rate
35.7%
74 of 207 trades profitable
74 Wins
133 Losses
Wins avg +20.00% eachLosses avg −4.30% each
Positive Expectancy
Profitable despite <50% win rate
Top Performers (selected trades)
TICKER
HOLD
P&L %
EXIT
GE
638d
+141.96%
Stop Hit
META
174d
+67.32%
Stop Hit
MU
169d
+61.28%
Stop Hit
CDNS
422d
+48.40%
Stop Hit
META
203d
+44.81%
Stop Hit
SCHW
410d
+74.41%
Stop Hit
CEG
189d
+34.05%
Stop Hit
TMUS
230d
+33.38%
Stop Hit
Sample trades from 207-trade ledger Request full CSV
The 2022 Bear Market Gauntlet

When the market lost a third of its value, the Sentry didn't.

2022 was the true test. Inflation, rate hikes, and the tech wreckage took QQQ down −32.6%. The S&P 500 lost −18.2%.

The Sentry's cash-shield engaged during high-VIX regimes. New entries were blocked. Trailing stops exited positions before the worst of the drawdown. The portfolio ended 2022 at −13.7% — far less damage than either benchmark. Read the post-mortem →

That capital preservation is why 2023–24 compounded so effectively. You can't grow what you've already lost.

The compounding math

A QQQ holder who lost −32.6% in 2022 needed a +49.7% gain just to recover. StockSentry's −13.7% required only +16.6% — capital that stayed in play for the +35%+ years that followed.

2022 Return Comparison
StockSentry
−13.7%
Capital preserved: $8,580 of $10,000 intact
SPY
−18.2%
−4.4pp worse than StockSentry
QQQ (NASDAQ 100)
−32.6%
−19pp worse · needed +49.7% just to recover

Max drawdown intra-year peak-to-trough: QQQ hit −46% at its worst. StockSentry: −23.7% at its worst (full 4-year period).

Methodology

How we backtested.

Point-in-time replay on daily OHLCV data. Every signal generated only from data available at that moment — no look-ahead, no in-sample peeking, no optimization on this period.

SP500 + NASDAQ100 universe (only high-quality tickers)
0.05% per-leg slippage on every fill
Agent guided trailing stop and stop-loss
Rotation delta rule: incoming conf > outgoing
VIX-regime bear blocking on new entries

Data

Daily OHLCV · Selected high-quality tickers from S&P500 and NASDAQ · Jan 2021–Dec 2024 · 1,007 trading days

Capital & Sizing

$10,000 starting equity. Up to 10 concurrent positions. 10% allocation per new entry; scale-in tranches at 10%.

Exit Logic

StockSentry's own exit logic based on ATR, 200-SMA and investor's personal.

Entry Logic

StockSentry's own logic based on trend persistence, sector regime, golden cross, SMA and volume. Read the 3-Prong Rubric deep-dive →

No Look-Ahead

Rubric frozen before the audit ran. No parameters touched the 2021–2024 window during tuning. Strict point-in-time replay.

Costs Modelled

0.05% per-leg slippage · $0 commission · No overnight margin · No dividend reinvestment. Conservative for StockSentry-identified quality S&P and NASDAQ names.

Honest Questions

The skeptic's FAQ.

We believe in radical transparency. Ask anything.

Is this real money or a simulation?

Simulation — the same rule-set the live bot uses today, replayed on historical data. StockSentry does not execute trades; it generates signals you act on with your own broker. This page reports what that signal stream would have produced under the stated assumptions.

Did you tune the model on this period?

We tuned the model across various regimes randomly to find out which technical parameters are the best indicative of a profitable position. However, the model was finalized before this audit run. The model had a strick no forward-looking context, and the results are purely done using the tecknical data on the data of testing and prior to that.

Why does the agent have only a 35.7% win rate?

By design. The system is optimized for asymmetric payoff, not accuracy. Stops exit losers fast (avg −4.3%) while winners are held until trailing stops are hit (avg +20%). A 35.7% win rate with a 4.65× average win/loss ratio produces strongly positive expectancy. Chasing a higher win rate typically sacrifices the big winners.

What's the −23.66% max drawdown about?

This is a ⚠️ flagged breach of an internal 15% guideline. It occurred during the 2022 bear regime. Context: QQQ hit −46% peak-to-trough in the same period. The cash-shield limited exposure but couldn't eliminate it entirely when entered positions turned quickly.

Will QQQ repeat its +55.9% 2023 and beat StockSentry again?

It might. 2023 was an exceptional AI/tech bounce year. StockSentry still returned +36.7% in 2023 — beating SPY by +10pp. The strategic advantage is in bear markets: while QQQ can lose −33% in a bad year, the Sentry's downside protection means you're never trying to dig out of a catastrophic hole.

Can I see every trade?

Yes. All 207 closed trades — entry/exit dates, tickers, prices, P&L, exit reasons — are available on request. Email alpha@stocksentry.bot. We log everything: the winners, the ugly −16% SMCI stop, all of it.

+122.8% tested · Beats SPY · Beats QQQ

Let the machine do the math.

StockSentry does the scanning, the stop-loss math, the risk/reward calculation, the exit timing — 24/7. You get a notification only when it's time to act.

View All Plans
Free to Apply
14-Day Pro Trial
No Auto-Subscription

Disclaimer. StockSentry is an algorithmic data tool, not a registered investment adviser. Backtest v9 results reflect simulated trading on historical data and do not represent actual trading results. Live performance will vary based on execution, timing, and market conditions. Past performance is not predictive of future results. Trading involves real risk of loss — never trade with capital you cannot afford to lose. Seer Labs Inc., Ontario, Canada.